Glastonbury Festival & Temporary Event Mapping

Communities United Kingdom
1

For several years, the Glastonbury Festival site has had a large number of temporary paths and other features permanently mapped:

image
image1544×1442 78.5 KB

image
image1544×1058 114 KB

These are not present on the ground, except during the event, and may move position year-to-year. The only way to know this for sure pre-event is to copy from Glastonbury’s copyrighted map.

This is also well-established bad OSM mapping practice, and although some of these features have been lifecycle-tagged for some of their time, this has not been consistent.

After discussing this in the IRC channel a few months ago, and with this year’s event having finished and the next Glastonbury not for two years, I have now deleted all the temporary infrastructure from OSM, along with doing a broader cleanup of the area (changeset #168440080). Hopefully we will get some aerial imagery next year which shows the farm in its default state.

I have kept the field names for the moment - they are arguably copyrighted too, and sometimes don’t match up with actual field boundaries, but some of them probably do match up with what those fields are actually known by during the rest of the year.

Anyone trying to do mapping or routing for festivals really should store this data outside of OSM, in a repository which is less strict about copyright status (I really don’t think Glastonbury will care, but that’s not good enough for OSM). It’s something I’ve been playing with and meaning to finish for years, but ironically I now work on mapping for Glastonbury Festival itself. I am happy to lend technical advice however!

6 Likes
2

Hey Russ, I’m one of the mappers guilty of such practice. I recently spent time at the site updating details, doing what I thought was reasonable. I think you’ve maybe been a little zealous in removing the paths that you have done, but I have no way to know if what looks like an established track from the imagery and being on the ground is actually used throughout the year on the farm, so you’ve probably taken the right action. As you’re working with the festival, maybe you can ask for details like the field names to be shared. It seems to me it would be rare for a farm to have copyrighted field names, but Worthy Farm isn’t any other farm.

Looking in the iD background layers and ‘Esri World Imagery (Clarity) Beta’ seems to show the farm in non-festival mode. I suspect the covid years, perhaps. Some of the tracks you’ve removed are visible.

3

This is extremely disappointing.

It is not really correct to suggest that paths of a festival which for at least 20 years has had the same layout (its evolutionary changes not withstanding) from year to year are merely ‘temporary’. These paths go well beyond that standard, namely ‘customary’ or at the very least ‘informal’.

This is not a temporary event in the sense of say a two-day festival in Brockwell Park. It is a site that is much-loved by hundreds of thousands of people, with a layout that is familiar from year to year.

The paths drawn in were based on extremely strong personal knowledge of annual experience on the site, not copied from any map, and represent customary paths used on a regular basis (almost-annually) by literally tens/hundreds of thousands of people. The marked informal routes provide actual utility and interest, namely those looking at the map between the festival periods to understand the layout of the land and provide memories of how they moved about. They enable people to geolocate photographs properly also. Those visiting the farm outside the festival period (with the consent of the landowner, which is well-known to be more liberal than many land owners) also would arguably welcome the customary routes being shown.

You’ve also been over-zealous in their removal. Even if you do not accept the customary paths, there are clearly permanent paths removed, as shown from the ESRI Clarity imagery. For instance, Creep Ground to King’s Meadow (‘Green Futures’ to the ‘Stone Circle’ fields) is wiped out. Even over Pennard Hill, the imagery shows clear evidence of trails.

Significant permanent land features have also been removed, e.g. the long drops, or for instance the bridge path parallel to Bella’s Bridge (a few metres east). These are permanent features of the landscape and there should be no dispute on them.

We took care never to add stages, pop-up infrastructure, glamping locations, or food/retail outlets: only customary routes through the landscape.

The only way to know this for sure pre-event is to copy from Glastonbury’s copyrighted map.

This is not a relevant point. Lots of things in OSM around the world are modified shortly after appearing on the ground rather than before. (In any case, those present on the site from the Sunday when general setup moves in could perform updates.) OSM is not providing some kind of guarantee of correctness for the event - that is not the purpose of micro-mapping.

Burning Man is the nearest equivalent case, and that remains mapped in heavy detail. Almost certainly that is because, like Glastonbury, it is mapping that is genuinely of interest to people outside the festival period.

We would ask that the changes be reverted. There is no strong reason to remove perfectly useful data on routes customarily taken.

Arguably the correct compromise would be for the changes to be reverted, and for any paths not visible on the land to be set as informal=yes if they were not already present.

PS On a more positive note, the festival app’s use of mapping this year is a significant enhancement over the previous generation app: congratulations to whoever did that.

1 Like
4

On the point of this reference:

https://wiki.openstreetmap.org/wiki/Good_practice#Don’t_map_temporary_events_and_temporary_features

That is aimed at dealing with the situation of routing systems being stuck with incorrect data - for instance due to a street closed for an event causing bad apps that don’t update regularly then avoiding the area for months. However, the situation is entirely different here:

a) People will not get routed through the site anyway as there is no external connectivity and in any case, routes should be marked access=private, so there is no detriment; and

b) As noted above, these customary/informal routes are of genuine interest even after the event has gone. Their limited presence on the ground for around 9 months of the year does not change this.

5

I think we should have community consensus to keep this data in OpenStreetMap, particularly as we are at the start of a fallow period for Glastonbury, with the next event at the end of June 2027.

I’m happy to revert my own changes if that is the consensus of the OSM GB community. If the decision is not to revert, I will certainly fix the handful of permanent paths I erroneously deleted – I didn’t realise the Esri Clarity imagery was different from classic Esri.

(I did make a lot of landuse improvements at the same time; that’s my own damn fault, I am happy to lose them.)

If all these paths were to be restored, I would at least want to see them without the name tags (which seem to be somewhat arbitrary), and likely with an access=customers tag.

Regarding the Burning Man comparison, I note that the layout in that case is copied from their official maps – I can only assume this is with permission (1, 2). Curiously the wiki seems to claim that it isn’t mapped in OSM any more and this note appears to suggest that discussion isn’t yet settled.

At any rate, Black Rock City is quite literally in the middle of nowhere, and you’re unlikely to find yourself accidentally routed over those roads. Glastonbury is much closer to civilisation. I’d argue more comparable festivals are Roskilde (which admittedly does have more genuinely permanent infrastructure) or Wacken.

Ironically, these show the pre-2020 layout, which is no longer current, while the paths which I deleted were different (and correct for the most recent layout). At any rate, those are clearly ghosts of paths and I wouldn’t really bank on them being easily walkable:

image
image1802×1742 419 KB

The above-ground superstructure of the long drop toilets is a temporary structure which is removed between events, and they are then capped to leave a flat surface. They are, perhaps, man_made=storage_tank, location=underground for 95% of the year. Leaving them tagged as amenity=toilet, as they were, risks misleading walkers in need of the loo!

That accessible footbridge is also a temporary structure and is only present for the event.

Someone certainly did - Arcadia was on there, along with a number of camping areas tagged as campsites which I also removed.

I can’t take credit for that, as my mapping work was for staff only. I think you have Vodafone, or most likely their subcontractor, to thank. It was pretty good! Maybe one day they’ll work out how to do routing on that app…

Honestly, my main objection here is that you’re only ever going to be able to do a mediocre job of mapping Glastonbury in OSM. The toilets which were mapped are a tiny subset of the number of toilets; there’s so much stuff which you can’t add which would make it a much better map.

I don’t really see how mapping these temporary features significantly helps people attending Glastonbury or those who happen to be around the site for the rest of the year.

6

Thanks. Some quickfire replies due to other things tonight:

  • The new land use stuff changes would be valuable to keep. (This is an argument for more atomic commits of course…). It’s also great the field names are in place - we avoided copying them and you clearly have superior knowledge.
  • Agree names on Ways could be dropped (as a compromise, though they help orientation).
  • The point about Burning Man was not about it being copied but rather that it is data left in place between the festivals.
  • Routing would be avoided with access=customers or whatever. If that wasn’t there before it should have been.
  • Agreed Pennard Hill is changed, and pretty sure we did an update reflecting that probably about 2023 when it shifted.
  • Thanks for the correction about the toilets and Bella’s Bridge (the latter is surprising given that the bridge itself has an attribution/story plate visible only when that side bridge is present). Agree storage tank tagging seems correct. It would be useful for these not to be lost. One set in the woods has been lost, the other to the south of the woods was missing and we were about to fill it in today! They are useful orientation-wise, and they avoid people wondering what on earth these weird landscape things are from the aerial imagery! But our main focus was on the walkways as they really help show how the festival fits together geographically.
  • Yep, definitely campsite locations shouldn’t be there.
  • On the very first point re community consensus to keep, our key point is that we don’t see these as temporary routes in the sense the wiki reference is about, and that they are long-established informal/de-facto routes people will recognise, and have genuine utility particularly after the event, particularly for those of a geographical bent. So for that reason these don’t initially have a clear reason to omit.
7

As far as I know, multiple years of Black Rock City have been mapped and deleted not long afterwards. The challenge with this major recurring event is that, if you delete it, someone will eventually see it’s missing and restore it. Leaving it in OSM is ironic given Burning Man’s emphasis on ephemerality.

OpenHistoricalMap has a couple years of Black Rock City mapped. If we can get more years in OHM, then I think it would be a more compelling alternative, so we don’t have to have this back and forth every year after the event. We could leave a few demolished:place=* note=* features around in OSM to point post-Burners to OHM.

8

To add to the evidence base, a similar scenario exists at the location for the Boomtown festival near Winchester. I’ve reached out to a seemingly single-purpose OSM-user who has mapped the festival grounds using irregular or incorrect tags temporary=* & frequency=*:

The user stated that the source for these changes is the festival-produced map (see original changeset from 2019):

Might there possibly be a place for including map-able features for regularly re-occuring events such as festivals in the OSM database? Would, of course, need to meet the normal OSM criteria, i.e: publicly verifiable (not only by customers) & appropriately licensed. Perhaps this would allow renderers / data users to determine whether to show the data on their maps by querying for a suitable check_date=* ?

9

The challenge with including “something that is only there for a couple of days a year” to OSM is that many or most data consumers don’t update from OSM daily - in some cases its technically difficult or actually impossible for them to do so.

“Whether to map a temporary layout in OSM” is always a bit of a judgement call. It’s similar to the questions about road closures - if something’s closed for a couple of months, it probably belongs in OSM, but a couple of days, probably not.

That said, it’s perfectly possible to use OSM-based tools to show “temporary” data over the top of OSM - uMap is one obvious option, as is a simple Leaflet overlay of data that is “similar to, but does not belong in OSM”. If anyone would like to do this but doesn’t know how to, please ask!

I usually go to a few music festivals a years and (apart from infrastructure that’s there year-round) haven’t added stage and path layouts etc. despite the Bing imagery of that location being from about 11:00 from the first day of the festival!

10

For what it’s worth, OpenHistoricalMap is also an option for time-limited data that would otherwise belong in OSM if it were more permanent. Currently the area is blank other than the former railway, but in principle someone could add the Glastonbury Festival details there and the renderer will take care of the rest.

Technical details

OHM doesn’t have a superb solution for features that recur exactly in the same place. Each recurrence would need to be a distinct overlapping feature, since there’s currently no support for a “repeat every [interval]” tag. In practice, we usually find it necessary to make new features anyways. For example, this annual Christmas tree display needs to have different features due to varying species=* tags and information about where each tree came from, while Burning Man’s Black Rock City moves around every year (e.g., 2008 versus 2009). That said, if someone were to come up with a good case for a recurring feature tag, we could get the developers to add support for it somehow.